Update README.md

README.md

@@ -22,51 +22,104 @@ size_categories:
 **Author:** Neerav Gupta  
 **Code:** [github.com/Neerav-Gupta/tokamark-robustness](https://github.com/Neerav-Gupta/tokamark-robustness)
 
+---
+
 ## Dataset Description
 
-This dataset contains pre-processed numpy arrays and trained model checkpoints from the first systematic robustness benchmark of plasma diagnostic ML models under realistic sensor failure, using the [TokaMark benchmark](https://arxiv.org/abs/2602.10132) on MAST tokamak data.
+This dataset contains pre-processed numpy arrays, trained model checkpoints, and experiment results from the first systematic robustness benchmark of plasma diagnostic ML models under realistic sensor failure, using the [TokaMark benchmark](https://arxiv.org/abs/2602.10132) on MAST tokamak data.
+
+We evaluate four architectures (XGBoost, LSTM, Transformer, and the TokaMark CNN baseline) across six physically-motivated failure scenarios and three imputation strategies, and compute shot-level alarm metrics using ground-truth disruption timestamps from FAIR-MAST.
+
+The raw data originates from the [FAIR-MAST dataset](https://github.com/UKAEA-IBM-STFC-Fusion-FMs/tokamark) provided by UKAEA, IBM Research, and STFC. This dataset contains derived arrays prepared for robustness benchmarking and is not a redistribution of the raw FAIR-MAST data.
+
+---
+
+## Key Results
+
+| Model | Clean NRMSE | Robustness Score (RS) | Clean TPR | TPR (proximate 25%, zero-fill) |
+|---|---|---|---|---|
+| XGBoost | 0.494 | 0.841 | 0.40 | — |
+| LSTM | 0.496 | 0.808 | 0.52 | 0.00 |
+| Transformer | 0.470 | 0.765 | 0.48 | 0.08 |
+| CNN (TokaMark baseline) | 0.528 | 0.764 | 0.60 | 0.46 |
 
-The raw data originates from the [FAIR-MAST dataset](https://github.com/UKAEA-IBM-STFC-Fusion-FMs/tokamark) provided by UKAEA, IBM Research, and STFC. This dataset contains derived arrays prepared for robustness benchmarking experiments and is not a redistribution of the raw FAIR-MAST data.
+**Key finding:** Under disruption-proximate sensor failure, LSTM alarm detection collapses to TPR = 0.00 with zero-fill imputation, but recovers to TPR = 1.00 with mean-fill — the opposite of its effect on NRMSE.
+
+---
 
 ## Dataset Structure
 
 ```
 tokamark-robustness-data/
-├── train_X_feat.npy
-├── train_X_ts.npy
-├── train_y.npy
-├── val_X_feat.npy
-├── val_X_ts.npy
-├── val_y.npy
-├── test_X_feat.npy
-├── test_X_ts.npy
-├── test_y.npy
-├── test_raw_samples.pkl
-├── feature_names.json
-└── checkpoints/
-    ├── xgboost_clean.pkl
-    ├── lstm_clean.pt
-    └── transformer_clean.pt
+├── data/
+│   ├── train_X_feat.npy          # XGBoost feature vectors, train
+│   ├── train_X_ts.npy            # Time series tensors, train
+│   ├── train_y.npy               # Target labels, train
+│   ├── val_X_feat.npy            # XGBoost feature vectors, val
+│   ├── val_X_ts.npy              # Time series tensors, val
+│   ├── val_y.npy                 # Target labels, val
+│   ├── test_X_feat.npy           # XGBoost feature vectors, test
+│   ├── test_X_ts.npy             # Time series tensors, test
+│   ├── test_y.npy                # Target labels, test
+│   ├── test_raw_samples.pkl      # Raw test samples with t_cut timestamps
+│   └── feature_names.json        # Feature names for X_feat columns
+├── checkpoints/
+│   ├── xgboost_clean.pkl         # Trained XGBoost model
+│   ├── lstm_clean.pt             # Trained LSTM model
+│   ├── transformer_clean.pt      # Trained Transformer model
+│   └── cnn_clean.pt              # Trained CNN baseline model
+└── results/
+    ├── xgboost_results.json
+    ├── lstm_results.json
+    ├── transformer_results.json
+    ├── cnn_results.json
+    ├── shot_level_metrics.json
+    ├── alarm_under_corruption.json
+    └── alarm_mitigation_proximate.json
 ```
 
+---
+
 ## File Descriptions
 
+### Data Arrays
+
 | File | Description | Shape |
 |---|---|---|
 | `train_X_feat.npy` | Training feature vectors for XGBoost | (9950, 142) |
-| `train_X_ts.npy` | Training time series for LSTM/Transformer | (9950, 600, 18) |
-| `train_y.npy` | Training targets | (9950,) |
+| `train_X_ts.npy` | Training time series for LSTM/Transformer/CNN | (9950, 600, 18) |
+| `train_y.npy` | Training targets (normalized plasma current) | (9950,) |
 | `val_X_feat.npy` | Validation feature vectors | (2500, 142) |
 | `val_X_ts.npy` | Validation time series | (2500, 600, 18) |
 | `val_y.npy` | Validation targets | (2500,) |
 | `test_X_feat.npy` | Test feature vectors | (2420, 142) |
 | `test_X_ts.npy` | Test time series | (2420, 600, 18) |
 | `test_y.npy` | Test targets | (2420,) |
-| `test_raw_samples.pkl` | Raw test samples for corruption experiments | 2420 samples |
-| `feature_names.json` | Feature names for X_feat columns | 142 names |
-| `checkpoints/xgboost_clean.pkl` | Trained XGBoost model | — |
-| `checkpoints/lstm_clean.pt` | Trained LSTM model | — |
-| `checkpoints/transformer_clean.pt` | Trained Transformer model | — |
+| `test_raw_samples.pkl` | Raw test samples including `t_cut` disruption timestamps and signal time arrays | 2420 samples |
+| `feature_names.json` | Feature names for the 142 X_feat columns | 142 names |
+
+### Checkpoints
+
+| File | Description |
+|---|---|
+| `checkpoints/xgboost_clean.pkl` | Trained XGBoost model (clean data) |
+| `checkpoints/lstm_clean.pt` | Trained LSTM model (clean data) |
+| `checkpoints/transformer_clean.pt` | Trained Transformer model (clean data) |
+| `checkpoints/cnn_clean.pt` | Trained TokaMark CNN baseline (clean data) |
+
+### Results
+
+| File | Description |
+|---|---|
+| `results/xgboost_results.json` | Full robustness results for XGBoost |
+| `results/lstm_results.json` | Full robustness results for LSTM |
+| `results/transformer_results.json` | Full robustness results for Transformer |
+| `results/cnn_results.json` | Full robustness results for CNN baseline |
+| `results/shot_level_metrics.json` | Clean-data shot-level TPR and MWT |
+| `results/alarm_under_corruption.json` | Shot-level alarm metrics under sensor failure |
+| `results/alarm_mitigation_proximate.json` | Alarm metrics under proximate failure with each imputation strategy |
+
+---
 
 ## Data Details
 
@@ -76,7 +129,11 @@ tokamark-robustness-data/
 | Val | 50 | 2,500 |
 | Test | 50 | 2,420 |
 
-**Task:** Task 4-4 from TokaMark — plasma current quench prediction. Given 150ms of diagnostic history across 14 input signals and 4 actuator signals, predict plasma current 100ms into the future.
+**Task:** Task 4-4 from TokaMark — plasma current quench prediction. Given 150ms of diagnostic history across 14 input signals and 4 actuator signals (18 total channels), predict plasma current 100ms into the future.
+
+**All 50 test shots disrupted** (finite `t_cut` in `test_raw_samples.pkl`), enabling real shot-level alarm metric computation.
+
+---
 
 ## Signal Order in X_ts (18 channels)
 
@@ -101,6 +158,8 @@ tokamark-robustness-data/
 | 16 | pulse_schedule-n_e_line | Actuator |
 | 17 | summary-power_nbi | Actuator |
 
+---
+
 ## Loading the Data
 
 ```python
@@ -116,18 +175,22 @@ snapshot_download(
 )
 
 # Load arrays
-X_train_ts = np.load("./tokamark_robustness_data/train_X_ts.npy")
-y_train    = np.load("./tokamark_robustness_data/train_y.npy")
-X_test_ts  = np.load("./tokamark_robustness_data/test_X_ts.npy")
-y_test     = np.load("./tokamark_robustness_data/test_y.npy")
-
-# Load raw test samples for corruption experiments
-with open("./tokamark_robustness_data/test_raw_samples.pkl", "rb") as f:
-    test_samples = pickle.load(f)
+X_train_ts = np.load("./tokamark_robustness_data/data/train_X_ts.npy")
+y_train    = np.load("./tokamark_robustness_data/data/train_y.npy")
+X_test_ts  = np.load("./tokamark_robustness_data/data/test_X_ts.npy")
+y_test     = np.load("./tokamark_robustness_data/data/test_y.npy")
 
 print(f"Train: {X_train_ts.shape}")  # (9950, 600, 18)
 print(f"Test:  {X_test_ts.shape}")   # (2420, 600, 18)
 
+# Load raw test samples (includes t_cut disruption timestamps)
+with open("./tokamark_robustness_data/data/test_raw_samples.pkl", "rb") as f:
+    test_samples = pickle.load(f)
+
+# Each sample has: shot_id, window_index, input, actuator, output, t_cut
+print(f"Sample keys: {list(test_samples[0].keys())}")
+print(f"t_cut (disruption time): {test_samples[0]['t_cut']:.4f}s")
+
 # Load trained LSTM checkpoint
 import torch
 ckpt = torch.load(
@@ -135,8 +198,19 @@ ckpt = torch.load(
     map_location="cpu"
 )
 print(f"LSTM n_features: {ckpt['n_features']}")
+
+# Load trained CNN checkpoint
+ckpt_cnn = torch.load(
+    "./tokamark_robustness_data/checkpoints/cnn_clean.pt",
+    map_location="cpu"
+)
+print(f"CNN n_channels: {ckpt_cnn['n_channels']}, "
+      f"input_len: {ckpt_cnn['input_len']}, "
+      f"backbone_hidden: {ckpt_cnn['backbone_hidden']}")
 ```
 
+---
+
 ## Reproducing Results
 
 ```bash
@@ -154,13 +228,21 @@ snapshot_download(
 )
 "
 
-# Run experiments
+# Train all four models and run robustness evaluation
 python scripts/train_xgboost.py
 python scripts/train_lstm.py
 python scripts/train_transformer.py
+python scripts/train_cnn_baseline.py
+
+# Compute shot-level alarm metrics
+python scripts/compute_alarm_metrics.py
+
+# Generate all 9 figures
 python scripts/analyze_results.py
 ```
 
+---
+
 ## Citation
 
 If you use this dataset please cite:
@@ -184,7 +266,7 @@ Please also cite the original TokaMark benchmark:
              Plasma Models},
   author  = {Rousseau, C{\'e}cile and Jackson, Samuel and
              Ordonez-Hurtado, Rodrigo H. and Amorisco, Nicola C. and
-             Boschi, Tobia and Holt, George K and Loreti, Andrea and
+             Boschi, Tobia and Holt, George K. and Loreti, Andrea and
              Sz{\'e}kely, Eszter and Whittle, Alexander and
              Agnello, Adriano and Pamela, Stanislas and
              Pascale, Alessandra and Akers, Robert and
@@ -195,10 +277,14 @@ Please also cite the original TokaMark benchmark:
 }
 ```
 
+---
+
 ## License
 
 [MIT License](https://opensource.org/licenses/MIT)
 
+---
+
 ## Acknowledgements
 
 Raw plasma data sourced from the FAIR-MAST dataset provided by UKAEA, IBM Research, and STFC. This benchmark dataset was prepared independently using the TokaMark data loading infrastructure.